PUBLISHED IN IRSE NEWS | ISSUE 231 | MaRcH 2017 1INTERNaTIONaL TEcHNIcaL cOMMITTEE TOPIc 40

aspects of producing a business case for ERTMSPrepared on behalf of the IRSE International Technical Committee by Rod Muttram

The purpose of the IRSE’s International Technical Committee (ITC) is to provide thought leadership and disseminate learning on technical topics relevant to train control and communication systems. This provides value not only to IRSE members but also to the wider rail industry. The committee’s particular strength lies in its international membership, enabling engineering principles and practices from a diverse range of countries to be brought to bear upon the subjects that are debated.

In this report Rod Muttram explains the findings of the committee’s investigation into the business case for ERTMS, particularly in light of recent major projects.

IntroductionThere have been many attempts to produce a business case to justify the implementation of ERTMS/ETCS either in a particular area/country or on a particular line of route. With a few notable exceptions such as Denmark and Norway many of them have ended in perceived ‘failure’ because the incremental benefits delivered relative to any alternative approaches are not seen as sufficient to justify the incremental costs. As a result there have been a lot of negative comments that ERTMS/ETCS is the ‘wrong system’ and that railways should either continue with conventional line-side signalling or wait for the next generation of technology to emerge.

The ITC disagree. To a large degree these issues arise because of the way in which ERTMS/ETCS has evolved and the additional costs that have resulted; now it is time to move on.

ERTMS/ETCS has been shaped as an interoperable system and thus the ‘political’ justification and its underpinning economics have always been ‘supranational’ and several business cases have been produced that show positive benefits at that level.

The need to align the system to the requirements of all European Union (EU) member states (the number of which has grown throughout the specification development) to facilitate internal as well as cross border use has delayed finalisation of the specifications by many years and led to many iterations of the system development. In this climate any business case at the individual country or route level was and is bound to be challenging. But the latest Memorandum of Understanding (MoU) signed between the European Commission, the European Union Agency for Rail (ERA), the rail sector associations (representing the Infrastructure Managers and Railway Undertakings) and the signalling supply industry irse.info/ot9h1 commits to stability in the Specifications at Baseline 3 Maintenance Release 2 (BL3 MR2) and to make all existing and future wayside installations interoperable with a Baseline 3 on-board. So whilst the specifications may still not be seen by some as ‘perfect for all uses’ a distinct point of stability has been reached.

ERTMS/ETCS is now really the ‘only show in town’ and it is time to ‘up the pace’ and drive costs down to produce the necessary economic and other benefits at the national and route level.

current contextIn the last 20 years or so rail has enjoyed a renaissance in many areas. Many new cities have constructed Metros, and existing systems have been added to and/or extended. High speed rail has been a huge success in France, Germany, Spain, and Italy. Eurostar and Thalys have provided popular and successful international services in competition with short-haul airlines and in recent years China has opened over 19000 km of new lines since the first commenced service in April 2007; most of these lines use the CTCS-3 version of the Chinese Train Control System which has ETCS Level 2 at its core so that many times more ETCS kilometres have been run in China than in the rest of the world.

The UK has seen significant growth since privatisation in the early ‘90’s driven by a number of factors and on the inter-city and commuter routes is facing a capacity crisis. In the short term government has had to resort to the same unpopular measure used by governments of all colours for at least 40 years and choke off demand with high fares (or at least not subsidise to keep them down). Creating new capacity is a lengthy and expensive process, the ‘Digital Railway’ initiative using improved train control and supervision must be a key part of creating that capacity in a more affordable and timely way; and timeliness is becoming more imperative as the graph in Figure 1 below, courtesy of the European Union Agency for Rail, shows. At the European level all the efforts since 1995 establishing the Common Safety Method (CSM), Interoperability Directives and Technical Specifications for Interoperability (TSI) have only slowed the decline in freight and despite all the successes overall passenger market share has only returned to its 1995 level, and the growth is showing signs of levelling off.

Further, rail is also facing new challenges. Particularly within Europe, reductions in the volume of the traditional bulk freight items such as coal and steel will be hard to replace; and if rail cannot address its high capital and operational costs, and the capacity issues in the areas where growth could occur, then new, more agile and less constrained solutions such as autonomous vehicles and road freight vehicle ‘fleeting’ are likely to not only meet the growth needs but to erode rail’s core market share further.

For those of us who believe in rail as an energy efficient and low land footprint mode of transport that would be a tragedy; none of us want to return to the days of decline.

So, in the ITC’s view the time for discussion on the way forward or to wait for some new ‘better’ or ‘lower cost’ solution are over. Whilst it is accepted that deployment of ETCS on a mixed traffic railway is more complex than high speed, the Chinese experience shows what can be done in a very short time with the right focus and determination. Railways need to start seeking solutions rather than constantly looking for problems, and that includes being willing to change operational rules and practices rather than always looking for the technology to cover every eventuality and solve every problem. The ETCS specifications have been through three major iterations largely to accommodate the huge range of diverse national requirements across EU Member State

International Technical committeeReport on topic 40

PUBLISHED IN IRSE NEWS | ISSUE 231 | MaRcH 20172 INTERNaTIONaL TEcHNIcaL cOMMITTEE TOPIc 40

railways. One fundamental enabler to reducing the cost is to now keep the specifications stable so that the suppliers can amortise their accumulated development costs over an increasing pool of sales rather continuing to rack up new costs. This is particularly true for the on-board system where software development and homologation costs far exceed the cost of the hardware. There is also an urgent need to build up a pool of skilled resources for critical activities such as rolling stock fitment. That needs a well-planned and consistent deployment, not ‘stop-go’. In that regard the MoU detailed above is considered a very positive step.

It is recognised that the national and supranational factors will still make it necessary for the EU and national governments to provide ‘pump priming’ funding. In the words of the MoU preamble para VIII “The timely realisation of the ambitious goals of ERTMS deployment will depend on the availability and efficient use of financial support at national and EU level.”

So now is the time to get on with ETCS deployment but to enable that some key factors need to be controlled to enable the maximum benefits at the minimum cost. Again, the MoU recognises this at preamble para X and tasks the ERA with coordinating efforts: “The Agency should drive a roadmap on ERTMS maturity and cost reduction, supporting the sector in producing tender templates, software evolution clauses and ensuring that the Control Command and Signalling TSI revision process will not cause uncertainty for users’ investments”.

This paper will attempt to set out some of factors to be considered in getting to an optimum scheme design and implementation strategy, based on the Committee’s experience, with the aim of helping to make the best case in all situations.

The factors to be considered in any ETCS investment/business case are considered in turn. They are not necessarily in priority order as that will depend on the specific circumstances of the area or line of route under consideration and the rolling stock that uses it. Nor are they necessarily comprehensive as there may be local factors and circumstances that bring in additional considerations (whilst aiming not to customise the equipment or its software beyond the specific application).

Factors for considerationPolitical/standards complianceETCS is mandated in EU member States for all new mainline railway routes and all existing routes on upgrade or substantial renewal unless a specific derogation is sought. There is, therefore, an argument that says any Business Case is irrelevant; it is a matter of legal compliance and we should only be seeking the most cost effective implementation matched to the needs of the route. It is as yet unclear what the UK vote to leave the EU will mean in this regard, but making the assumption that Network Rail will move wholly outside this framework in the short term seems foolish in the extreme. Outside of the EU Directive compliance is clearly not required, and ETCS has to stand on its merits against other options. It is often compared in terms of costs and performance to ‘Metro’ type Communications Based Train Control (CBTC) systems. The complexity of the ETCS System Requirement Specification (SRS), which provides functionality to be able to implement many different railways’ operating rules and principles, is one of the main drivers of the system cost and thus far has largely negated any competitive cost advantages that might have come from having a standard solution from many manufacturers. The SRS has evolved over many years with three major revisions each of which has largely changed the details rather than the core functionality but has required the main functional software to be re-written and re-homologated.

Some believe that to save costs the SRS needs to be simplified (but who will give their functions up); others favour a sub-division into a set of simpler ‘class or route type’ specifications, but that is likely to be politically unacceptable as it defeats the whole purpose of interoperability (even if in many cases such interoperability will never be exercised in practice).

Further, with such complex software, re-homologation times are long and further delays are really the last thing needed. The manufacturers are tired of change and believe that there is now a need for a period of stability to allow the substantial ‘one off costs’ already incurred to be recovered. Certainly many implementation strategies require substantial numbers of trains to be fitted with the on-board system before ETCS can be commissioned and any benefits delivered and, as stated in the

0%

5%

10%

15%

20%

25%

30%

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Train passenger-km as % of (rail+cars+bus)

Train tonne-km as % of (rail+road+waterways) ERA establishedTSIs - 1st wave

Interoperability Directive full coverage

TSIs - 2nd wave

TSIs -3rd wave

CSMs

Rail freight market opening

Passenger Freight

Figure 1 – Share of rail traffic volumes in Europe since 1995. Source ERA.

PUBLISHED IN IRSE NEWS | ISSUE 231 | MaRcH 2017 3INTERNaTIONaL TEcHNIcaL cOMMITTEE TOPIc 40

introduction, software development and homologation costs amortised over too few unit sales are the main driver for the current high cost of on-board systems. The constant changes have also been a major cause of unreliability as some of the more obscure software bugs are only found after many hours and/or km of field operation. Stability of the specifications going forward should also enable real reliability growth to be achieved. CTCS-3 in China has achieved greater than 1 million kilometres between service affecting failures so it can be done, but all software changes inevitably introduce new bugs. The need to recover the substantial investment (of company money) made in ETCS Level 1 and 2 is also a significant barrier to further Level 3 development.

So, the industry should now ‘run’ with Baseline 3 MR2.

All trains not ordered with ETCS already fitted should as a minimum be ‘ETCS ready’ with space and key interfaces provided ready for ETCS installation.

Any re-signalling in areas where ETCS is not a current requirement should also attempt, as far as possible, to install ‘ETCS ready’ interlockings (see the ITC paper on ETCS Level 2 over relay interlockings in the October 2016 IRSE NEWS). What ‘ETCS ready’ means is also subject to some interpretation and there is merit in considering defining this more substantially including the possible provision of ‘locked out’ or ‘bridged out’ ETCS features within the interlocking software.

Change in Headway/Capacity/Reliability This is often cited as a major reason for and benefit of ETCS deployment. However, as well as the capability of the chosen ETCS system itself, this will also depend on other limiting factors on the route(s) concerned and any improvement will be relative to the capability of the existing system being replaced, which if it has been ‘tuned’ and ‘optimised’ in the past can be quite high.

Factors such as flat junctions, point swing times, power supply capacity etc. may all need to be addressed as well if additional capacity is to be released. A comprehensive systems evaluation is essential at the design stage and these additional costs will affect any business case (including any alternative control system to ETCS if it is to achieve the same system level outputs).

Of course there is also no point in providing capacity that is not needed and despite the pressure on some parts of the infrastructure many routes remain underutilised. Justifying ETCS

on these routes just for interoperability or technical commonality will always be difficult viewed in isolation.

Whole Life Cost Whole life cost (WLC) will be built from a wide range of factors and assumptions including the cost of the equipment and its projected maintenance (which may require significant skill-base changes) and operating costs over the planned lifecycle all compared to the equivalent costs for a range of alternatives including life extension of the existing systems. Technology lifecycles are becoming shorter and architectures which allow hardware to be updated whilst preserving key safety critical software assets are to be commended.

Estimates must include all costs including things like the need for additional power supplies, both for the signalling and the traction if more trains are to be run on an electrified route. It is also very important to consider timing issues (when will spend occur relative to when any benefits will start to be delivered?) and organisational factors (in a railway involving multiple entities in delivery, where do costs fall, where do benefits fall and what are the mechanisms for aligning these where they are not naturally aligned?). This is where governments and Regulators have a key role to play, both in mandating system fitment, aligning benefits and where necessary providing bridging funding where costs precede benefits to a significant degree.

Network Rail in the UK and Prorail in the Netherlands are currently working together on a migration strategy that they are calling ‘Hybrid’ ETCS. This involves overlaying ETCS Level 3 (L3) over existing wayside infrastructure and then allowing L3 equipped trains with a suitable Train Integrity Monitoring System (TIMS) to run in L3 at tighter headways than the fixed block structure would normally permit using either moving block or virtual block within the physical track circuit or axle counter sections. Trains with ETCS but no TIMS can run in Level 2. Unequipped or cab equipment/radio failed trains can either be run using procedural controls (with the control centre still able to track them) or on lineside signals if those are retained. This would allow transition to L3 to take place without 100% train fitment with both ETCS and TIMS. Lower track access charges for fully equipped trains that can run at the reduced headways would create an incentive for operators to fit whilst occasional users of the track section concerned can delay until the other routes over which they run are fitted.

Successful deployment of ETCS requires widespread fitment of ETCS on-board equipment. Taking the UK case, ETCS comes fitted ‘as standard’ on stock including (from left to right), Siemens Class 700 for Thameslink, Bombardier Class 345 Crossrail fleet and the Hitachi Class 800/801 for GWR and Virgin. Photos Siemens, Bombardier, Hitachi.

PUBLISHED IN IRSE NEWS | ISSUE 231 | MaRcH 20174 INTERNaTIONaL TEcHNIcaL cOMMITTEE TOPIc 40

Eventually when 100% fit is achieved the wayside infrastructure can be de-commissioned. Similar architectures have already been used for transition/migration in the Metro domain. Estimation of the whole life cost of this architecture requires judgements to be made about how long the residual track based train detection will need to be retained and maintained and may involve starting re-signalling earlier than would perhaps normally be the case, when the wayside elements are considered to still have the required remaining residual life, rather than reaching ‘life-expired’ condition as would more normally be the case. Book value for the updated system may comprise assets of several different ages at different points in their depreciation cycle. (For more detail on this architecture see “A Recipe for a fast Nationwide Migration to Robust ERTMS L3” by Martin Zweers and Maarten Bartholomeus, IRSE NEWS Issue 193, October 2013 and similar articles in SIGNAL+DRAHT, September 2013 and Railway Gazette International, September 2015).

It is also important not to forget ‘soft’ factors like the need to train and maintain the competency of staff in the new technologies and operating rules, which can also be substantial and logistically challenging.

SafetyETCS is considered by the ITC to be the current ‘state of the art’ in terms of railway control systems for mainline railways. The incremental safety benefit brought by ETCS for existing lines will be highly dependent on the performance of the signalling or automatic train protection (ATP) system it replaces.

The reduction in line-side infrastructure offered by ETCS Level 3 or the Hybrid L2/L3 architectures currently being explored may also bring significant reductions in track worker risk.

British Rail’s 1994 analysis of ATP showed that the benefit it brought over and above that delivered by the then standard Automatic Warning System (AWS) system was insufficient to justify its cost on safety grounds. It was deemed to be ‘not reasonably practicable’ against benchmark values of preventing fatality. The UK now has supplemented AWS with the Train Protection and Warning System (TPWS) and most European railways have similar ‘Class B’ legacy systems. When compared to the safety levels delivered by these legacy systems the incremental safety benefit from ETCS over their maintenance/renewal will thus be very small and is thus unlikely to form a significant part of any business case justification.

There will really only be a significant safety benefit where the railway or route concerned has no or little stop aspect or over-speed enforcement. That said, the ITC is concerned that in the 21st century there are still significant railway routes which lack consistent and fully effective over-speed enforcement as the recent Santiago de Compostela derailment in Spain, the Spyten Dyvel derailment on Metro North in the US (now being fitted with Positive Train Control - PTC - the US ATP solution) and the TGV derailment at Strasbourg in France all serve to illustrate in different ways. Further, the collision at Bad Aibling demonstrates yet again that whilst it is operationally necessary to be able to manually override a stop signal the processes for doing that must be extremely robust. All these examples yet again demonstrate the importance of Human factors (and thus error proofing) in overall safety system design (a subject on which the ITC plans to produce an article soon). Older or disparate systems with multiple interfaces are more prone to these kinds of issues and a

great deal of human factors work has gone into the ETCS design. The ITC question whether it really remains acceptable in the 21st century, with the technology that now exists, for any significant railway to continue to operate relying on the driver alone for speed control or a signaller/dispatcher alone to override a stop signal without decision support. ETCS and PTC are both better considered as essential modern safety systems and part of the current ‘state of the art’ preferred solution. It is the proposal to maintain or renew any alternative system which delivers a lower level of safety that should (in the ITC’s view) have to be fully justified.

Equipment availability For most manufacturers ERTMS/ETCS is their current mainline ATP product with little else available as a modern offering. In the ITC’s view PTC is optimised for a predominantly heavy haul freight railway and its use of on-board equipment without a formal Safety Integrity Level (SIL) would be unlikely to be acceptable to most railways operating predominantly passenger services even where ETCS is not mandatory. In Europe the choice is therefore largely between ETCS and extending the life of heritage systems. For non-European Union routes with simple characteristics a metro derived CBTC solution is also a possibility. These alternatives often introduce significant uncertainties into life cost calculations. However, one of the significant arguments against a standard like ETCS is that it stifles innovation and the fact that specific actions are having to be taken to sustain the GSM-R train radio system as the commercial mobile communications world moves from 4G to 5G emphasises this so issues such as obsolescence need to be carefully considered for each specific scheme.

In a world where commercial technology is moving so fast rail needs to think differently. The critical investment now is in software not hardware and it is that that needs to be ‘future proofed’ by keeping the specification stable and making it as ‘platform independent’ as possible.

conclusionThe problem faced by any ETCS business case is that both the costs and the benefits are influenced by a wide range of the above listed factors, some generic and others location, route, train or organisation specific. This impacts to such an extent that the synthesis of any ‘overall’ case becomes a very complex exercise the outcome of which is dependent on the judgemental selection of many options and assumptions and which is also impacted by a wide range of uncertainties. Any case produced is therefore relatively easy to challenge simply by selecting some different, but probably equally supportable, options.

The situation is not helped by the baseline from which any business case is judged being highly variable.

In the ITC’s view there is no ‘do nothing’ option if we want to have a safe, efficient and competitive railway system going forward and the time for agonising over business cases for ETCS is over. The way to drive the cost of ETCS down is to get on with deploying significant volumes of a consistent standard across many railways and stop ‘paralysis by analysis’.